Atomically Precise Manufacturing — Part 2

A look at the methods available today for manipulating individual atoms. And how we might one day grow this into the ultimate manufacturing method.

In Part 1 of this article we talked about what APM is, and explored a couple of the most likely applications for this technology as it becomes more feasible. In this segment we will discuss what is possible today with our current technologies, and the major hurdles that must be overcome for this to become a practical means of production.

The methods we have today for manipulating individual atoms are still a far cry from the universal assemblers that Eric Drexler envisioned in his seminal 1986 book Engines of Creation: the coming era of nanotechnology. Drexler imagined that we would be using artificial nanoscale molecular machines, a synthetic analogy of the protein-based molecular machines that are prevalent in biological systems, capable of assembling any object imaginable atom by atom. But although we have not yet reached this capability we have made noteworthy progress in this direction in the form of nanofabrication, synthetic biology, self assembly, and scanning probe manipulation techniques.

Nanofabrication / Nanomanufacturing

Before discussing how APM can be accomplished, it is important to understand a few terms:

• Nanomanufacturing is described by the National Nanotechnology Initiative (NNI), a US Government R&D Initiative, as: “Manufacturing at the nanoscale… Nanomanufacturing involves scaled-up, reliable, and cost-effective manufacturing of nanoscale materials, structures, devices, and systems”
• With nanoscale being “structures with a length scale applicable to nanotechnology, usually cited as 1–100 nanometers” (Wikipedia)
• Top-down: “Top-down fabrication reduces large pieces of materials all the way down to the nanoscale, like someone carving a model airplane out of a block of wood” (NNI)
• Bottom-up “creates products by building them up from atomic- and molecular-scale components” (NNI)
• Nanofabrication is often used interchangeably with nanomanufacturing. However, nanofabrication is used more to convey an R&D activity consuming funds, while nanomanufacturing is a production activity generating revenue (Nanostructure Control of Materials)

Currently in most industries top-down manufacturing is the dominant form, with examples being machining, lumber, mining, and textiles/garments, where large pieces are reduced to smaller shapes. This can be performed at the micro and nano-scale with etching processes, or laser nanomachining. Etching is the process of removing a material from the surface of another material. Wet etching, also known as chemical etching involves chemical reactions to remove material from a sample. Dry etching, or plasma etching removes material by exposing a surface to high energy particles. Laser nanomachining is the more sophisticated method in which a surface is exposed to high energy photons with very short wavelengths or very short femtosecond pulses to remove material with nano-scale precision on the surface of a sample. Both laser machining and plasma etching are used in the semiconductor industry to create features a few nanometers in size on silicon substrates.

Bottom-up manufacturing is closer to the end goal of APM, assembling macroscale materials from individual atoms or molecules. But bottom-up manufacturing is still in its infancy. A few examples of bottom-up manufacturing that we might all be familiar with are additive manufacturing (3D printing) and chemical synthesis (the construction of complex chemical compounds from simpler ones, and a common process in biological organisms).

Top-down still remains the dominant form today even when we consider nanomanufacturing, largely because of limitations on our ability to manufacture things at the atomic and molecular levels. However, top-down is far more wasteful because it “requires larger amounts of materials and can lead to waste if excess material is discarded” (NNI). Bottom-up manufacturing would use only the material that is needed to create the end product, meaning there is no excess being wasted. Right now, bottom-up processes are constrained by the time it takes to utilize them, a lack of ability to operate precisely at the nanoscale, or both. By gaining the ability to operate at this scale, material waste could be eliminated or at least dramatically reduced.

Scanning Probe Manipulation Techniques:

Scanning probe manipulation techniques are a bottom-up way of implementing APM by building a structure one atom at a time. These techniques are derived from the scanning tunneling microscope (STM), which was invented in 1981 and is used to image conductive samples at an atomic level. The STM is used inside an ultra-high vacuum and uses a conductive tip with an electron on the end of it to scan the surface of the sample electron by electron. As the STM goes through this process, the electrons of the sample and tip tunnel through the gap between them; this tunneling allows the STM to create an image of the scanned surface. In 1994, Joe Lyding, a professor at the University of Illinois Urbana-Champaign, first demonstrated that an STM could be used to inject electrons into the chemical bonds of a silicon surface, enabling individual hydrogen atoms to be removed and replaced with other compounds that bonded with the hydrogen-terminated silicon atoms in advantageous ways. This technique is known as hydrogen depassivation lithography (HDL), and it shows how an STM can be used to remove and add atomic particles and can also be used to build up a structure atom by atom. Although this method is not used in commercial and industrial scale production because it is painstakingly slow and difficult to scale, 2019 saw what Zyvex Labs (a leading R&D player in the APM space) called “the first practical application of HDL” come out of the University of New South Wales (UNSW):

Researchers created nanoscale electronic devices, including wires, transistors and quantum dots (e.g. interconnects and um-scale bond pads) to connect to the outside world. The HDL patterns were filled with phosphine and, using a short anneal, the phosphorus atoms were incorporated as dopants into the silicon, thus creating metallic conductors, electrodes, gates, and even the ‘single-atom transistor,’ where one isolated phosphorus dopant forms the channel of the transistor.” (Zyvex Labs)

The use of HDL and STM in concert reduces the imprecision inherent in other processes for injecting atoms into surfaces. Atomically-precise tools like this will be necessary to scale quantum computers from a single qubit to many qubits, and they will need to be both reliable and high-throughput. In order to get there, companies like Zyvex are working to use AI to automate the process. In April 2020, Zyvex received a $1m award from the Department of Energy to work on automation of HDL using AI algorithms to both automate the lithography process and identify errors. This would allow for us to “remove the tedious, time-consuming manual processes [of STM and HDL] to enable a fully autonomous atomic precision lithography process. This will improve the productivity and reliability of research tools and pave the way for highly parallel tools that could be used for manufacturing of products such as quantum computers” (SBIR.gov).

In addition to the work out of the UNSW and Zyvex, there are other recent developments that could help to make scanning probe manipulation techniques more viable. In March 2019, Dr. Reza Moheimani, a systems engineering professor at UT Dallas, received a $2.4m grant from the Department of Energy for his use of STM for high-throughput APM. The STM in this research essentially functions as a robotic arm that picks hydrogen atoms off of a sample of silicon and then replaces them with other types of atoms, building the surface up layer-by layer with precision and without defect. In terms of scaling this process, UTD Today provides a good high-level outline of this: “Moheimani’s group is developing microelectromechanical systems (MEMS) actuators, each a fraction of an inch in size and equipped with an atomically sharp tip that could function as a high-speed STM. These MEMS STMs can be scaled up in large arrays to enable high-throughput lithography for atomically precise manufacturing of novel materials and quantum electronic devices.” In the same UTD Today article, David Forrest, of the Advanced Manufacturing Office at the DOE, touched on the throughput breakthroughs made: “The traditional speed of operation of these scanning probes currently prevents meaningful production rates. However, Professor Moheimani’s innovations break through long standing barriers, with the potential for a thousandfold increase in speed and parallelization.” The result of this work could eventually be a way to build products atom by atom without defect at scale, thus achieving a future with APM.

There are a number of obstacles standing in the way of this future that we must find ways to overcome. In order to perform the process at scale, thousands of STM tips with the ability to move in unison in different directions and with nanometer precision would be critical. Additionally, right now a conductive sample is needed for the process to work and the STM has to operate in an ultra-high vacuum, often at a cryogenic temperature. Also, widening the size and scope of APM using this process is challenging, since as Moheimani notes: “we don’t have to solve just one problem; we have to solve a dozen challenging problems and develop new systems and feedback control methods. Then we must put them together to get them to work simultaneously.”

While STM and HDL methodologies of implementing APM are viable and exciting, their applicability is limited to surfaces that conduct, like metals or silicon used in the semiconductor industry. This gets at an idea discussed earlier: the semiconductor industry may not be able to drive any more innovation in APM/nanomanufacturing for other industries; other industries need tools that work on non-conductive surfaces.

Chemical Vapor Deposition

Chemical Vapor Deposition (CVD) is a mature manufacturing technique that relies on self assembly to grow structures or uniform coatings on a substrate. These coatings and structures are grown by introducing vaporized material into a special chamber, which encourages the atoms in the vapor to leave the gaseous phase and deposit onto a surface in an ordered manner. While this wouldn’t count as APM since we are not controlling where individual atoms fall, we have learned all manner of clever techniques to get atoms to *mostly* behave as intended; so the result could rarely be considered defect-free. One of CVDs intriguing applications is in the creation of carbon nanotubes (CNTs); these nanotubes are “cylindrical molecules that consist of rolled-up sheets of single-layer carbon atoms (graphene)…. [and are] ultra-high strength, low-weight materials that possess highly conductive electrical and thermal properties” (Nanowerk, Carbon Nanotubes). CNTs have use cases in automotive manufacturing, energy storage, actuators, water filters, and much more. CVD is a process in which a material is placed inside a vacuum chamber and a coating is vaporized using heat and then introduced into the chamber, where it settles on the material in a uniform manner. The process of using CVD to create CNTs is described well in Synthesis of Carbon Nanotubes by Catalytic Chemical Vapor Deposition: “In the process of CVD synthesis of CNTs, water vapor, oxygen (air), and ethanol are added into the growth atmosphere, and they can selectively remove amorphous carbon without damaging the nanotube and significantly improve the activity and life of the catalyst.” The creation of the CNT itself and the coating of the CNT both rely on different CVD processes — one to create the nanotube out of layers of carbon atoms and another to coat the nanotube; MIT mechanical engineering professor John Hart observes that “by combining two CVD processes… we have a scalable way to manufacture nanomaterials with new properties” (MIT). CVD allows us to create CNTs with the properties that make them so useful, and illustrates the power that self-assembly processes have in allowing us to create stronger, more advantageous materials.

Synthetic Biology and pharma assembly lines

Today one of our leading methods for making new molecules with specific geometries is through synthetic biology, in which we mutate the genome of bacteria to hijack their biology and trick them into assembling interesting compounds for us. Much of this is done by trial and error, doing random mutations and then keeping the ones that express something useful. This could be thought of as a form of APM. However this would just form the building blocks or feedstock from which larger macro-scale objects could be assembled (or self-assembled). If we can achieve a more digital approach to APM the implications for medicine would be profound. Imagine if we could simply assemble these novel organic compounds directly, perhaps with an army of nano-machines that build the molecules one atom at a time. It might not make it any easier to know what a particular compound is good for, whether poison or medicine, but it would surely make it a lot easier to reproduce the ones that do prove to be beneficial.

Of course, it does raise interesting questions about the bad stuff too. This will be a platform by which bioweapons could be produced as well. And likely a fair number of such hazardous compounds will be accidentally discovered by researchers pursuing therapeutics and medicine. But any technology can be used for harm by bad actors. For now we can trust that these advances will lead to more good than harm, as has been the case with almost all technologies that mankind has harnessed.

Stay Tuned

In Part 3 of this article we will explore in more detail some of the challenges holding this branch of research back, along with some promising solutions. And we will also take a closer look at how to image and inspect your work at these nanometer scales to determine if you have achieved atomic precision.

Prime Movers Lab invests in breakthrough scientific startups founded by Prime Movers, the inventors who transform billions of lives. We invest in companies reinventing energy, transportation, infrastructure, manufacturing, human augmentation, and agriculture.

Sign up here if you are not already subscribed to our blog.